24–29 April 1993 | Jakob Nielsen and Thomas K. Landauer
The paper by Jakob Nielsen and Thomas K. Landauer presents a mathematical model to predict the number of usability problems found in user testing and heuristic evaluation. The model is based on the assumption that the detection of usability problems follows a Poisson process, where the probability of finding a problem is independent of previous tests. The authors analyze data from 11 studies and find that the model fits well, with high R² values. The model can be used to estimate the number of evaluations needed to achieve desired levels of thoroughness or benefits, and to determine the optimal number of evaluators or test users based on cost-benefit analysis. The results suggest that for a medium-sized project, 16 evaluations would be cost-effective, with the maximum benefit-to-cost ratio at four evaluations. The model also accounts for the variability in problem detection rates among evaluators and test users, and can be applied to iterative design processes. The authors conclude that the Poisson model is a useful tool for planning and evaluating usability efforts in software development.The paper by Jakob Nielsen and Thomas K. Landauer presents a mathematical model to predict the number of usability problems found in user testing and heuristic evaluation. The model is based on the assumption that the detection of usability problems follows a Poisson process, where the probability of finding a problem is independent of previous tests. The authors analyze data from 11 studies and find that the model fits well, with high R² values. The model can be used to estimate the number of evaluations needed to achieve desired levels of thoroughness or benefits, and to determine the optimal number of evaluators or test users based on cost-benefit analysis. The results suggest that for a medium-sized project, 16 evaluations would be cost-effective, with the maximum benefit-to-cost ratio at four evaluations. The model also accounts for the variability in problem detection rates among evaluators and test users, and can be applied to iterative design processes. The authors conclude that the Poisson model is a useful tool for planning and evaluating usability efforts in software development.